Targeted agents are revolutionizing cancer treatment. However, important challenges remain. In particular, even among patients with the same known mutation that sensitizes them to a particular targeted therapy, there is a significant range of responses to treatment, from no response (progressive disease) to complete response (e100% tumor volume reduction). What drives this response variability is poorly understood, and response to treatment is generally determined after the fact. In addition, tumors invariably develop resistance to treatment and recur. Identifying-early in the course of therapy-patients that will or will not respond to a given therapeutic regimen and predicting the durability of response would be of enormous clinical benefit: In addition to limiting patients' exposure to the toxicities associated with unsuccessful therapies, it would allow patients the opportunity to switch to a potentially more efficacious treatment. As there are many therapeutic regimens available, and many more being developed, switching treatment early in the course of therapy is a very real option-but only if a reliable method to determine early response were available. Unfortunately, existing methods of determining response and progression are inadequate, as they require long clinical observation times with consequent discomfort, financial burden as well as inability to pursue alternative options. The overall goal of this project is to integrate quantitative in vitro and in vivo imaging measurements to predict the maximum patient tumor response early in the course of oncogene-targeted therapy, in order to enable alternative treatment options that minimize or prevent the emergence of the resistant phenotype. A major barrier to this goal is the lack of quantitative data dynamically linking clinical tumor response t underlying response at the cellular level. Preliminary studies show the feasibility of combining imaging modalities at three biological scales: 2D culture, where drug response can be quantified accurately and dynamically by automated microscopy; 3D bioreactor, more closely simulating in vivo and addressable both by microscopy and magnetic resonance (MR) imaging; rat brain tumor xenografts, an excellent preclinical drug treatment model suitable to MR imaging. The three levels will be integrated by mathematical models incorporating quantifiable parameters and suitable to in vivo validation. In Aim 1 we will optimize extraction of parameters from 2D and 3D microscopy and MR imaging data of the erlotinib-responsive (PC9-DS9) and resistant (PC9-BR1) human lung cancer cell lines, well-studied models for oncogene-addicted lung cancer. From these data we will establish a look up table of proliferation and death rates linking 2D microscopy and 3D bioreactor MR estimates. In Aim 2 we will quantify tumor growth dynamics of erlotinib-treated DS9/BR1 mixed cultures in the 3D bioreactor, by initializing and constraining an image-based model. In Aim 3 we will test predicting acute resistance to oncogene directed therapy in brain tumor xenografts of DS9/BR1 mixtures, by integrating in vivo MRI data with microscopy data and model them to monitor the spatiotemporal appearance of the resistant phenotype.